Project Overview
The goal of this project is to develop a detection system capable of capturing images of and identifying wildlife. The motivation is twofold: I recently witnessed a black fox (fairly rare) in my driveway but wasn't able to capture an image of it before it fled in order to identify it manually (until later), and additionally would like to potentially track when each of my dogs have been let outside for bathroom breaks as one needs to be let out more frequently. I plan to get some experience using AI analytics to aid in the identification, as well as bluetooth for the communications. Considering this is an individual project, will focus first on object detection, then object identification, and finally communication as time permits.
Buildroot
The primary hardware for this project was a Raspberry Pi Zero 2 W, chosen for its size, sufficient processing power for image recognition tasks, onboard wifi, a camera that's relatively easy to integrate, and buildroot hardware support. An imx219 Camera Module was be used for capturing images or videos. Data was be uploaded via the onboard wifi module to my Ubuntu machine. The hardware was all be sourced myself.
OpenCV
GStreamer
CMake
- Components of aesdsocket,
- modules,
- scripts,
- threads, and
- file IO
- Aspects of computer vision (OpenCV)
- Media network communication (GStreamer)
This got me through the biggest hurdle I encountered throughout this project.
Buildroot Repository hosted at https://github.com/cu-ecen-aeld/final-project-tiba6275/
Application code hosted in in a repository at https://github.com/tiba6275/ecen5713-final-project/
https://github.com/users/tiba6275/projects/2/views/1
I'm satisfied with where I ended up as I assumed I probably wouldn't make it to the AI portion of my original plan, but I would love to get to that eventually.
https://github.com/cu-ecen-aeld/final-project-tiba6275/wiki/Tim's-Final-Project-Video